About Axel
Skills: ROS2, Python, Rust, C++, Gazebo, Navigation (Nav2), Isaac Sim, PyTorch, OpenCV, Cuda, Docker, Linux, PostgreSQL, React, Blender, Unity, Behavior Trees, 3D perception, Nvidia Jetson, and general distributed systems.
I am a Robotics and Applied AI engineer with 6+ years of consulting experience. I build robotics and AI systems from architecture to deployment, have them installed onsite in the field with clients.
Past projects include ABB Saffron flower picker with IRB and FlexPicker robots, a deployed fleet of navigating autonomous golf-ball picking robots, drone-mounted hyperspectral sensing for landmine detection, reinforcement-learning flight route planning for Saab Aeronautics, and a high volume ML ATS matching system that improved candidate search for a recruiting firm.
I can help with:
- Robotics prototypes, ROS2 systems, perception, navigation, manipulation, and simulation
- Computer vision for messy sensor data, including segmentation, tracking, and 3D perception (SLAM, LiDAR)
- Isaac Sim, Gazebo, RobotStudio, Docker, synthetic data, and sim-to-real workflows
- Production ML systems with data pipelines, evaluation, vector search, and deployment
- Search, ranking, matching, and recommendation systems using LLMs, embeddings, PyTorch, Qdrant, PostgreSQL, and Redis
- AI automation tools, internal platforms, APIs, dashboards, and technical MVPs
- Architecture reviews for early-stage AI, robotics, and automation products
English
Native or bilingual
Swedish
Native or bilingual
Remote only
Primarily works remotely
Experience
- Dyno Robotics - Industrial Saffron Robotics SystemRobotic Software DeveloperOctober 2024 - June 2025 (8 months)Developed a robotic saffron sorting system that increased processing speed in a production plant, combining ABB industrial robots, ML-based computer vision, and custom motion planning in a live industrial pipeline.
- Built an automated manipulation and sorting workflow using ABB Robots
- Worked across ROS2, C++, Python, PyTorch, Docker, RAPID, and RobotStudio for deployment and industrial integration
- Integrated a Rust motion-planning controller with custom path planning for precise robot control
- Applied AI: RL and supervised learning for guiding manipulation, and integrated to allow for continuous live improvement
- Explored Gaussian Splatting and Blender-based integration as a separate research arm connected to the robotics stack
- Delivered the system as part of a real industrial production pipeline, with the constraints that come from uptime, safety, and operational reliability
- Dyno Robotics - Drone-Based Landmine DetectionRobotics Software DeveloperAugust 2024 - December 2024 (4 months)Software lead for R&D project on drone-based AI landmine detection using multi-spectral imaging, with work spanning applied machine learning, sensing pipelines, and decision support for human operators.
- Led the project and coordinated technical work around drone-mounted multi-spectral sensing for per-pixel landmine detection
- Explored supervised, unsupervised, and self-supervised learning approaches for detection and classification
- Worked with PyTorch-based vision models including UNet-style segmentation approaches and representation learning methods such as SimCLR
- Developed approaches for sensor fusion for the different color channels
- Worked on calibration and alignment for multi-camera setups
- Built the project around real-world constraints in remote sensing, noisy data, and limited ground truth
- Dyno Robotics - ATS Matching PlatformAI Software DeveloperJune 2023 - June 2025 (2 years)Led the architecture and development of a full-stack AI recruitment matching system for a recruiting firm, using custom neural ranking, vector retrieval, and integrated recruiter workflows to turn large-scale hiring data into better matching decisions.
- Owned the architecture and end-to-end development of a production ATS matching system used by a real recruiting firm
- Built the platform across Svelte, TypeScript, Python FastAPI, Rust, Docker, Postgres, Redis, Kafka, and Azure
- Implemented retrieval and ranking pipelines using Qdrant and a custom PyTorch neural network
- Trained reranking models continuously on live incoming recruitment data and recruiter feedback
- Built ETL workflows for large volumes of historical and incoming candidate and job data
- Integrated matching, tagging, evaluation, and workflow logic into recruiter-facing tools
- Designed the system around privacy, fairness, explainability, and high-volume exhaustive evaluation
- Helped give the firm a measurable quality and speed advantage over more manual recruiting workflows
Recommendations
Be the first to recommend Axel
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Exchange studiesEscola Politécnica da USP, São Paolo, Brazil2019Academic exchange focused on technology and engineering courses taught in Portuguese - Studied AI fundamentals, including mathematical foundations and Prolog programming - Worked with data science methods, including a supervised learning project in stock prediction - Completed coursework in virtual reality and game development using Unity3D
- BSc in Cognitive ScienceLinköping University2021-Interdisciplinary degree combining AI, neuroscience, psychology, scientific methods, and the study of human cognition and emotion - Built a strong foundation in both the technical and human sides of intelligent systems, with coursework spanning classical AI, neural networks, and language technology - Completed AI coursework covering classical methods and neural networks, including a project in simulated evolution - Worked on language technology, including a project applying a temporal t-SNE approach to Reddit data - Developed an interdisciplinary way of thinking that still shapes later work in robotics, machine learning, and human-centered system design